Dear all,

We are considering investing in the LabChip GXII instrument from Caliper/Perkin Elmer for automated analysis of protein samples (as a replacement for SDS-PAGE).  I already talked to Tim in Dortmund about his experience, but would be interested in hearing from other users as well.  We would share the instrument with a few other labs to get the necessary sample throughput to justify having the instrument.  For us the main advantages would be the automated quantification/percent purity calculations and ability to analyze a large number of samples quickly, plus the advantage of being able to prepare user reports more quickly (no scanning/labeling gels . . .).

What are other people's thoughts/experiences?

In addition we are also looking at the Janus BioTx pro system for running Atoll columns, also from Perkin Elmer, so if someone has experience with this system or Perkin Elmer robotics I would be interested to hear about that as well.

Thanks very much!



Posted on 19-Aug-2013 11:44 CEST
Peggy Stolt-Bergner

I have this instrument. We have stopped using it.

  • Chips are very expensive and clog up regularly
  • It is OK for visualising purifs, but complex mixes such as lysates or even low stringency NiNTA purifs are difficult to analyse.
  • Some proteins run at a completely different size to SDS-PAGE gels.
  • It's OK for analysing PCR products.

We worked really hard at getting it working, but finally everyone slipped back to running SDS-PAGE gels because they believed the results (especially faint bands and non-pure samples). The digitised virtual gels are really not very clear.

My view is that this would be fine for running large numbers of purified proteins (QC lab?), but has limited use in simply substituting for normal SDS-PAGE use.


Posted on 19-Aug-2013 11:53 CEST
Darren Hart

Hi Peggy,

We have one and had very similar experience as outlined by Darren.

DNA chip is fine but protein chip is not as robust or reliable. Initially we thought it's our technique or method but after speaking to others, we found that Tim at DPF is the only one who manages to use it regularly. We also worked very hard to try to get it going and even had their technical specialists worked alongside us. In the end, the MW inaccuracy and low sensitivity are not useful for screening.



Posted on 20-Aug-2013 0:43 CEST
Linda Lua

Dear Darren and Linda,

Thanks a lot for your feedback.  The instrument performed very well during the demo, but I'm glad to hear about your "real life" experiences.  We may have to reconsider.

I guess you have both gone back to traditional SDS-PAGE?  As far as I know there are no other such systems out there except the Qiaxcel from Qiagen, but we demo-ed that and it's the same - great for DNA, but poor performance for proteins.




Posted on 20-Aug-2013 8:36 CEST
Peggy Stolt-Bergner

Hi Peggy,


We have had a demo for the LabChip as well (about 3 years ago) for protein separation and although the results were quite OK, we decided not to purchase it because of logistical issues in assembling enough samples at the same time. Instead, we bought the TapeStation from a company called Lab901, which is now part of Agilent Technologies. They now sell it as TapeStation 2200 and if I am correct it can run both DNA, RNA and Protein samples on different chips. On our instrument, we can only run protein samples. The instrument uses chips containing 16 slots/capillaries. You don't have to use them all at the same time; unused slots can be used at a later stage. It takes about 10-15 minutes to prepare the samples (including two heating steps at 75 °) and the running time is about 12 minutes for a full chip. The result is shown either as a conventional SDS-PAGE or as a chromatogram and data can be processed: Automated quantification/percent purity calculations and easy preparation of user reports (no scanning/labeling gels).

Although we are quite satisfied with it, I have to mention that because of the costs (~2.5 - 3 euro per sample including reagents and chip) most people in the Structural Biology groups still use traditional SDS-PAGE to analyse multiple elution fractions and only sometimes use the TapeStation for final quality control. If you are interested, you might ask for a demo; at Lab901 that was easily arranged, I don't know how Agilent is organising that.

If you need more info, just let me know.



Posted on 20-Aug-2013 9:32 CEST
Patrick Celie

Hi all,

because our work is mainly focused on HTP protein expression screening on a wide number of proteins, benchmarking of fusions, strains…But also on developing micro-assays like  in vitro HTP protein-DNA interaction studies (HTP SELEX protocols are included in Jolma A et al ,Cell. 2013 Jan 17;152(1-2):327-39) and in vitro HTP protein-protein interaction studies (the HTP Hold-up article is submitted) our throughput at analytical scale is huge and we have purchased a GX II during the summer 2010 to try to cope with the sample numbers.

I lost track but we have run to date probably around 100.000 points in 96/384 format and gave more money for the GX II consumables to caliper than to any other company in the meantime...

We are overall very happy users of the machine that was the only way to achieve our goals in the timeframe. We use it mainly with soluble extracts or purified (Nickel 96 mainly) proteins. When we want to have a great result we use a new chip and a new kit for 384 samples (that is typically what we do when we run 384 format of the in vitro protein-protein assay with marginal error bars). But most of the time the precision that we need is not much more than the SDS_PAGE… For this kind of run, when we are cautious and when using pure proteins we typically use the same chip for >1500 points and up to 2000 points with some extra cleaning in between. With the soluble extracts we typically starts to have problems around 1000 proteins. We almost never clogged a chip, when we have dirty samples (total extracts, lysates) we filtrate them on Millipore 96 0,22 uM or we dilute them in SB with urea before filtration. We always boil (or reboil) the plates just before using them to be sure that the SDS or other stuff are not making particles and we freeze the plates afterwards in case or need to rerun later on. When the result are crap or when we clog one chip we sonicate the chip in water and usually we can use it again.

We probably saw all the problems that you can have on the machine over the years and if the first couple of years we really had very few problems then started the bad months when more than 50 % of our runs had problems and when we struggled and spent too much time/kit to try to fix them... Because of the heavy use of the machine our cleaning procedures were not good enough anymore and some traces of chemicals got stuck on the electrodes or elsewhere giving minor electricity/fluorescence/vacuum problems that were giving delays in time and errors in fluorescence/quantifications that were giving the failures. On top of that sometimes the chips (1 in 10 maybe) had manufacturing problems... At the end we took the service year warranty and since that everything is back to normal in most cases.

Again if you are cautious and if you need a bit of throughput I recommend the machine. If you don’t need to run on average more than 96 samples a week don’t bother and stick to SDS_PAGE… Because we needed the throughput we invested some time in the first place to have procedures that would extend the life span of the chip/kit, anyone can do this, nothing difficult (and we can help).

Some more informations that can be usefull:

-        Because we work with toxins for our European project (, which are very small (most are below 10 kDa), we started to use the High sensitivity chip (twice the price of the normal one), they are nice but much more sensitive to variation in buffer/chip age… So at the end the points gets very expensive, so avoid them as much as you can.

-        Yes compared to SDS_PAGE you have much more proteins that run at strange MW, maybe 20 % of the overall proteins that we run on it. But once you know it you are cautious and if needed after a Nickel 96 the 2-3 proteins that looked a bit odd are loaded on a gel to be sure.

-        When running 24 time the same sample you will have errors in the 10 % range or so in MW and quantification (remember than despite a gel that will have a smiling effect that are easy to spot on 15 samples here each sample run on a separate run) and that can get worst if you gel/chip is old. When we want to be precise (below 3 % on average) we add a internal protein in the samples to recalibrate MW/quantity exactly.

-        I find the machine quiet sensitive for our needs, we limit ourselves to detection limit of 5 ng/ul (that is the limit of the automated procedure) but on pure samples we go down to 1-3 ng/ul on manual mode without problems when needed.

-        People at caliper (and now PE) are much less experienced than the average people on this forum with proteins so don’t trust them too much and don’t expect too much from them when you have problems with the protein chips, we had most of the time to fix the things our own way (cleaning procedure…)

-        Since caliper is part of PE it is getting much more difficult when we run into trouble to get help, a chip replaced or a free kit because you run into troubles…


My 2 cents



Posted on 27-Aug-2013 16:25 CEST
Renaud Vincentelli

Hi Renaud,

Thanks a lot for your response.  I talked to some other LabChip users outside of the forum who are mainly analyzing pure proteins, and it seems for this it works fine, but as you also indicated problems can start with more complex samples.  They also told me that things have gotten more difficult after Perkin-Elmer took over.   After discussing it, right now we don't have the throughput for the LabChip, the idea was to share the instrument with several other labs, but the logistics will be difficult.

We are going to demo the TapeSation that Patrick is using next week, I will post the results here.

Thanks to everyone for your input!




Posted on 28-Aug-2013 12:41 CEST
Peggy Stolt-Bergner
Displaying 1-7 of 71